TR2023-046

Task-Directed Exploration in Continuous POMDPs for Robotic Manipulation of Articulated Objects


    •  Curtis, A., Kaelbling, L., Jain, S., "Task-Directed Exploration in Continuous POMDPs for Robotic Manipulation of Articulated Objects", IEEE International Conference on Robotics and Automation (ICRA), DOI: 10.1109/​ICRA48891.2023.10160306, May 2023, pp. 3721-3728.
      BibTeX TR2023-046 PDF
      • @inproceedings{Curtis2023may,
      • author = {Curtis, Aidan and Kaelbling, Leslie and Jain, Siddarth},
      • title = {Task-Directed Exploration in Continuous POMDPs for Robotic Manipulation of Articulated Objects},
      • booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
      • year = 2023,
      • pages = {3721--3728},
      • month = may,
      • publisher = {IEEE},
      • doi = {10.1109/ICRA48891.2023.10160306},
      • isbn = {979-8-3503-2365-8},
      • url = {https://www.merl.com/publications/TR2023-046}
      • }
  • MERL Contact:
  • Research Areas:

    Artificial Intelligence, Robotics

Abstract:

Representing and reasoning about uncertainty is crucial for autonomous agents acting in partially observable environments with noisy sensors. Partially observable Markov decision processes (POMDPs) serve as a general framework for representing problems in which uncertainty is an important factor. Online sample-based POMDP methods have emerged as efficient approaches to solving large POMDPs and have been shown to extend to continuous domains. However, these solutions struggle to find long-horizon plans in problems with significant uncertainty. Exploration heuristics can help guide planning, but many real-world settings contain significant task-irrelevant uncertainty that might distract from the task objective. In this paper, we propose STRUG, an online POMDP solver capable of handling domains that require long-horizon planning with significant task-relevant and task-irrelevant uncertainty. We demonstrate our solution on several temporally extended versions of toy POMDP problems as well as robotic manipulation of articulated objects using a neural perception frontend to construct a distribution of possible models. Our results show that STRUG outperforms the current sample- based online POMDP solvers on several tasks.

 

  • Related News & Events

  • Related Publication

  •  Curtis, A., Kaelbling, L., Jain, S., "Task-Directed Exploration in Continuous POMDPs for Robotic Manipulation of Articulated Objects", arXiv, December 2022.
    BibTeX arXiv
    • @article{Curtis2022dec,
    • author = {Curtis, Aidan and Kaelbling, Leslie and Jain, Siddarth},
    • title = {Task-Directed Exploration in Continuous POMDPs for Robotic Manipulation of Articulated Objects},
    • journal = {arXiv},
    • year = 2022,
    • month = dec,
    • url = {https://arxiv.org/abs/2212.04554}
    • }